TinyAstral: Transferring Self-Evolving Harnesses to Tiny Generalist Manipulation Agents
READ THE ORIGINAL SOURCEOpen original sourcehuggingface.co · Opens in a new tabTinyAstral is a two-stage approach for robot manipulation agents. An open-weight teacher's harness (tools, feedback, memory, recovery logic) is improved via recursive self-improvement, with harness edits accepted only if they pass a paired statistical test. The frozen harness is then used to distill small 0.8B-4B students that run in the same harness. On RoboTwin, teacher success rises from 33.4% to 63.6%, and the final 4B student reaches 62.8%. The student stays close to its teacher on six manipulation benchmarks at roughly one fifth of the accelerator time, and runs closed loop on a real arm.
One primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face Daily Papers
- Discovery linkHugging Face
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Open full article view →- Original paper+45
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
TinyAstral is a two-stage approach for robot manipulation agents. An open-weight teacher's harness (tools, feedback, memory, recovery logic) is improved via recursive self-improvement, with harness edits accepted only if they pass a paired statistical test. The frozen harness is then used to distill small 0.8B-4B students that run in the same harness. On RoboTwin, teacher success rises from 33.4% to 63.6%, and the final 4B student reaches 62.8%. The student stays close to its teacher on six manipulation benchmarks at roughly one fifth of the accelerator time, and runs closed loop on a real arm.
A READER'S EVIDENCE CHECKHow to read this signalSix questions about claims, sources, risks, and open checks
This is a signal-level guide until claim-level evidence is available. “Not recorded” means the evidence has not been assessed here; it does not mean a result is false. How we check AI news →
- 01Who is making each claim?
The claimant and whether each claim was measured or asserted are not recorded in this public brief.
- 02What was actually tested?
Not recorded per claim. Check the source for sample size, tasks, model and benchmark versions, and reproducibility.
- 03Where are the sources?
Original link: huggingface.co. Brief basis: fetched source page. Primary page date and quote locations are not established in this panel.
- 04What could mislead?
Source claims not independently checked; results are reported on specific benchmarks and measured deployment episodes.
- 05What is at risk if wrong?
Not assessed per claim. Consider safety, privacy, cost, and adoption decisions before relying on the result.
- 06What would change the view?
No claim-specific check is recorded. An inspectable primary method or independent replication could change the assessment.
Key findings from the stored brief
- Harness evolution raised the teacher's RoboTwin success from 33.4% to 63.6%.
- The final 4B student reaches 62.8% on RoboTwin, close to its teacher.
- Distilled students score close to their teachers across four harness versions, and later versions yield stronger students.
- On measured deployment episodes, the 4B student uses about one fifth of its teacher's accelerator time.
Source claims not independently checked; results are reported on specific benchmarks and measured deployment episodes.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
A publication revision history is not stored for this signal. New engagement snapshots do not mean the source itself changed.
Why this may matter
The work suggests that gains from improving an agent's surrounding harness can be carried into compact models, potentially lowering inference cost for embodied robot agents while keeping performance near the larger teacher.
Limits and open questions
Source claims not independently checked; results are reported on specific benchmarks and measured deployment episodes.
Organization-specific financial, security, and operational effects require the organization's own context and review.
Discussion
Comments are reviewed before publication. Discuss the evidence, not the person.
No published comments yet.
Reader discussion opens with accounts. You can read the original source and published comments now.